[PATCH 1/2] meta-selftest/staticids: Add dhcpcd

[PATCH 1/2] meta-selftest/staticids: Add dhcpcd

[Richard Purdie]

Add dhcpc's uid/gid to the static configs used by meta-selftest for
testing since dhcpc was added to OE-Core.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
 meta-selftest/files/static-group  | 2 +-
 meta-selftest/files/static-passwd | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/meta-selftest/files/static-group b/meta-selftest/files/static-group
index 9213b8e9bc0..da4dab53733 100644
--- a/meta-selftest/files/static-group
+++ b/meta-selftest/files/static-group
@@ -11,4 +11,4 @@ avahi:x:509:
 avahi-autoipd:x:510:
 rpc:x:511:
 rpcuser:x:513:
-
+dhcpcd:x:514:
diff --git a/meta-selftest/files/static-passwd b/meta-selftest/files/static-passwd
index 412f85d469f..6ce2886ceed 100644
--- a/meta-selftest/files/static-passwd
+++ b/meta-selftest/files/static-passwd
@@ -9,3 +9,4 @@ avahi-autoipd:x:510:510::/:/bin/nologin
 rpc:x:511:511::/:/bin/nologin
 distcc:x:512:nogroup::/:/bin/nologin
 rpcuser:x:513:513::/var/lib/nfs:/bin/nologin
+dhcpcd:x:514:514::/:/bin/nologin
-- 
2.27.0

[PATCH 2/2] selftest/reproducible: Add useradd-staticids to reproducible builds tests

[Richard Purdie]

Its surprising we've made it this far without this, add in some specific
uid/gid settings to ensure these don't cause false positives to the tests.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
 meta/lib/oeqa/selftest/cases/reproducible.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/meta/lib/oeqa/selftest/cases/reproducible.py b/meta/lib/oeqa/selftest/cases/reproducible.py
index e3597d5081a..eee494e5cc8 100644
--- a/meta/lib/oeqa/selftest/cases/reproducible.py
+++ b/meta/lib/oeqa/selftest/cases/reproducible.py
@@ -249,6 +249,10 @@ class ReproducibleTests(OESelftestTestCase):
             TMPDIR = "{tmpdir}"
             LICENSE_FLAGS_WHITELIST = "commercial"
             DISTRO_FEATURES_append = ' systemd pam'
+            USERADDEXTENSION = "useradd-staticids"
+            USERADD_ERROR_DYNAMIC = "skip"
+            USERADD_UID_TABLES += "files/static-passwd"
+            USERADD_GID_TABLES += "files/static-group"
             ''').format(package_classes=' '.join('package_%s' % c for c in self.package_classes),
                         tmpdir=tmpdir)
 
-- 
2.27.0

Re: [OE-core] [PATCH 2/2] selftest/reproducible: Add useradd-staticids to reproducible builds tests

[Alexander Kanavin]

[-- Attachment #1: Type: text/plain, Size: 1401 bytes --]

I wonder if this also addresses some of the existing entries on the
exception list?

Alex

On Sun 27. Dec 2020 at 13.10, Richard Purdie <
richard.purdie@linuxfoundation.org> wrote:

> Its surprising we've made it this far without this, add in some specific
> uid/gid settings to ensure these don't cause false positives to the tests.
>
> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> ---
>  meta/lib/oeqa/selftest/cases/reproducible.py | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/meta/lib/oeqa/selftest/cases/reproducible.py
> b/meta/lib/oeqa/selftest/cases/reproducible.py
> index e3597d5081a..eee494e5cc8 100644
> --- a/meta/lib/oeqa/selftest/cases/reproducible.py
> +++ b/meta/lib/oeqa/selftest/cases/reproducible.py
> @@ -249,6 +249,10 @@ class ReproducibleTests(OESelftestTestCase):
>              TMPDIR = "{tmpdir}"
>              LICENSE_FLAGS_WHITELIST = "commercial"
>              DISTRO_FEATURES_append = ' systemd pam'
> +            USERADDEXTENSION = "useradd-staticids"
> +            USERADD_ERROR_DYNAMIC = "skip"
> +            USERADD_UID_TABLES += "files/static-passwd"
> +            USERADD_GID_TABLES += "files/static-group"
>              ''').format(package_classes=' '.join('package_%s' % c for c
> in self.package_classes),
>                          tmpdir=tmpdir)
>
> --
> 2.27.0
>
>
> 
>
>

[-- Attachment #2: Type: text/html, Size: 2091 bytes --]

Re: [OE-core] [PATCH 2/2] selftest/reproducible: Add useradd-staticids to reproducible builds tests

[Richard Purdie]

On Sun, 2020-12-27 at 14:08 +0100, Alexander Kanavin wrote:
> I wonder if this also addresses some of the existing entries on the
> exception list?

I've wondered that too. We probably need to analyse the results of some
of the builds and see if any are reproducing that didn't previously...

We do have a couple of intermittent issues that are showing in world
builds, one in grub:

https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20201228-mcoi5qk0/packages/diff-html/

looking likely due to changes in:

/usr/src/debug/grub-efi/2.04-r0/grub-2.04/grub-core/unidata.c

and one in initramfs-base:

https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20201223-og6v1hn1/packages/diff-html/

where the write bits are masked out on /dev/console, something to do
with umask handling in pseudo on tumbleweed-ty-3.

I think one has a bug, the other does not (yet).

Cheers,

Richard

Re: [OE-core] [PATCH 2/2] selftest/reproducible: Add useradd-staticids to reproducible builds tests

[Naveen Saini]

> -----Original Message-----
> From: Richard Purdie <richard.purdie@linuxfoundation.org>
> Sent: Wednesday, December 30, 2020 10:40 PM
> To: Alexander Kanavin <alex.kanavin@gmail.com>
> Cc: openembedded-core@lists.openembedded.org; Saini, Naveen Kumar
> <naveen.kumar.saini@intel.com>
> Subject: Re: [OE-core] [PATCH 2/2] selftest/reproducible: Add useradd-
> staticids to reproducible builds tests
> 
> On Sun, 2020-12-27 at 14:08 +0100, Alexander Kanavin wrote:
> > I wonder if this also addresses some of the existing entries on the
> > exception list?
> 
> I've wondered that too. We probably need to analyse the results of some of
> the builds and see if any are reproducing that didn't previously...
> 
> We do have a couple of intermittent issues that are showing in world builds,
> one in grub:
> 
> https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20201228-
> mcoi5qk0/packages/diff-html/
> 
> looking likely due to changes in:
> 
> /usr/src/debug/grub-efi/2.04-r0/grub-2.04/grub-core/unidata.c
> 
> and one in initramfs-base:
> 
> https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20201223-
> og6v1hn1/packages/diff-html/
> 
> where the write bits are masked out on /dev/console, something to do with
> umask handling in pseudo on tumbleweed-ty-3.
> 
> I think one has a bug, the other does not (yet).
[SWAT] Already raised bugs for both.
grub : https://bugzilla.yoctoproject.org/show_bug.cgi?id=14167
Initramfs-boot: https://bugzilla.yoctoproject.org/show_bug.cgi?id=14162

Recently one more reproducibility failure observed with man-db package. 
Looks like Owner and Group of files '/usr/bin/man' and '/usr/bin/mandb' are missing/differnet.
https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20201230-cmrpjzcp/packages/diff-html/

Raised for this too.
https://bugzilla.yoctoproject.org/show_bug.cgi?id=14172

Already a bug#14159 exists for man-db-doc package. 

> 
> Cheers,
> 
> Richard